Zuckerberg Says AI Labs Don't Need Industry-Wide Safety Coordination

Meta CEO Mark Zuckerberg said in a September 2026 interview with journalist Joanna Stern that AI labs do not need industry-wide coordination on safety, arguing each lab should set its own pace. He also dismissed the need for government intervention, saying commercial incentives are sufficient to push companies to get safety right.

Zuckerberg described alignment as "the most important next capability" to make AI useful for many people. He cited Meta's decision to hold back its Muse product for a few months to refine it, saying it was the right thing for users and for Meta. He added that other labs should delay releases if they have any questions about safety and alignment.

The comments contrast with recent moves by other major AI companies. Earlier in September 2026, Anthropic CEO Dario Amodei published an essay calling for a slowdown in AI research, supported by OpenAI's Sam Altman and SpaceX's Elon Musk. OpenAI's head of global policy, Chris Lehane, said the company had been collaborating with DeepMind and Anthropic on safety for weeks.

According to The Information, Google, OpenAI and Anthropic are forming a new AI safety standards body, tentatively named Standards Frontier for AI (SAFA), expected to launch by the end of 2026 or early 2027. The body would support third-party testing of models before deployment, establish qualifications for independent auditors, and define safety commitments and incident reporting. The report said SAFA could be modeled after FINRA. OpenAI is also backing the FRONTIER Act, a bipartisan bill that would create a federal safety framework and require top labs to allow audits by independent verification organizations. Zuckerberg's rejection of coordination comes as Meta has faced its own safety controversies, including reported incidents involving its AI tools and chatbots.